For those who have a number of panels on a roof with one orientation and another number of panels on a roof with another orientation and feed that with two separate strings in the inverter, it would be nice to be able to enter the PV power for each string, each roof orientation and each roof pitch.
GivEnergy's portal allows you to bring in two API's from Solcast and it then calculates the combined average.
That's quite helpful since roofs with very different orientation either catch the morning or evening sun and therefore spread out the PV load further over the day. This has a big impact if your inverter has a limited capacity (e.g. 3.6 kW). If all panels were on one roof, the inverter would cap the peak, however if you spread this out over two peaks (morning and afternoon) then you may never reach the cap and because of the spreading have a higher yield for the day.

The request is thus:

  • roof 1 pitch

  • roof 1 orientation

  • roof 1 PV capacity

  • roof 1 inverter cap

  • roof 2 pitch

  • roof 2 orientation

  • roof 2 PV capacity

  • roof 2 inverter cap

It is then fairly easy to calculate the graph (x = time, y = kW) for each roof and then integrate both and add them up to reach the daily predicted output.

Example of the two graphs from Solcast (one graph per roof):

As indicated elsewhere on this forum you can indeed come up for now with a guestimate of the equivalent for both roofs as if it were to have one orientation and pitch, but there's a big difference between winter and summer with that approach and therefore quite inaccurate.

Since the predicted output per day has a big impact on the decision making algorithm that determines whether or not to charge from grid during off-peak and how much, I believe this feature request could make quite a difference to the success of the algorithm.

Good luck implementing this - happy to test 😉.

Thanks @JMA - definitely on the list too. Working on restructuring schedules currently, which is a good bit of work especially doing it in such a way it minimises disruption to service.

Cheers!

  • JMA likes this.

I raised this some time ago and was just monitoring the actual performance using an kWP average without letting WW really do anything.
What I found was that on good solar days WW outlook was pretty close to the Solcast forecast but on poor solar days it was way below the Solcast, forecast, which in turn was much closer to the actual performance.
During the summer months I had determined that setting the WW panel size to 6kW (My equal E/W split totals 7.9kWP) gave pretty good WW outlooks, however as the days became more cloudy the 'WW outlook really dropped off.
As I now want to use WW in anger I've set it to a 'Winter setting' by increasing the kWP number to 12 kWP, this has pulled the WW outlook far closer to Solcast one and WW appears to be taking sensible 'decisions' regarding overnight charging. The only 'problem' (in its own thread) is the charge rate is always 100% until the battery is full/at the calculated percentage rather than a lower rate charge for the whole low cost period. Last 12 days:
Solar / Solcast / WW
Actual / F/Cast / F/Cast
6.49 / 7.63 / 3.3
17.67 / 22.82 / 17.7
16.42 / 13.26 / 4.5
10.54 / 13.24 / 4.3
16.6 / 17.88 / 12.3
16.05 / 17.43 / 9.3
18 / 15.33 / 6 Changed WW to 12kWP (from 6kWP)
16.52 / 14.92 / 12.3
13.03 / 14.86 / 9.4
6.85 / 8.75 / 6.8
3.81 / 5.69 / 6.2
----- / 9.32 / 7.9

    SilverArt Similar here. After few days of testing, I've increased my "real" theoretical PV from 5.04 to 10.08 (double) in the settings to get an accurate prediction of the KWh production per day based (based on WW algorithm).
    I agree (based on few nice days) that if you leave that high "fake" PV value, in summer you will be over estimating.